Schone, Hunter Robert; (2023) Neurocognitive considerations for engineering bionic limbs. Doctoral thesis (Ph.D), UCL (University College London).

Preview

Text Schone_PhD_Thesis.pdf - Accepted Version Download (14MB) | Preview

Abstract

A longstanding engineering ambition has been to design anthropomorphic bionic limbs: devices that look like and are controlled in the same way as the biological body (biomimetic). Biomimetic design is built on the (untested) assumption that biomimetic devices might allow users to recruit pre-existing neural resources supporting the biological body to assist device control. In a series of studies, I investigated whether the human body is the ultimate design template for engineering bionic limbs. First, I explored how hands are represented in the brain, and how this representation changes after amputation. I found that, following amputation, the neural representation of the missing hand persists in primary somatosensory cortex. This evidence demonstrates the potential for neuroproshetic devices to harness these resources for control. As such, I next explored whether extensive use of a hand-replacement device (tool/prosthesis) would lead to a neural integration of representations for the device and the biological body (a process often referred to as embodiment). To accomplish this, I tested able-bodied expert tool users to understand whether they neurally represent their expert tool more like a biological hand. I found that expert tool users have more distinct hand and tool visual representations compared to novices, indicating no evidence for neural embodiment. Considering the lack of evidence for the neural embodiment of external devices, why then should we prioritize designing devices that mimic the human body? In the final study, I explored whether bionic limbs should be controlled in the same way as the human body (i.e., biomimetic control). I trained able-bodied participants to use a wearable myoelectric bionic hand. I compared motor learning across days and behavioural tasks for two training groups: Biomimetic (mimicking bionic hand gesture with biological hand) and Arbitrary (mapping an unrelated biological hand gesture with bionic gesture) control. I found that Biomimetic control was superior under a very specific set of circumstances, with Arbitrary control performing largely the same. Further, arbitrary users showed increased generalization to a novel control strategy. This work provides an important alternative framework to biomimetic design offering instead a neurocognitive overview of the challenges and opportunities for biomimetic and non-biomimetic control strategies. Collectively, these studies suggest that compelling opportunities exist for harnessing hand neurocognitive resources to control bionic limbs. Though, current technology allows for only limited integration.

Download activity - last month

Export as CSVXMLJSON

Download activity - last 12 months

Export as XMLCSVJSON

Downloads by country - last 12 months

Export as JSONXMLCSV

Archive Staff Only

View Item